ELCIC (version 0.2.1)

MLICwgee: Model selection based on MLIC under the syntax of WGEE (Main function).

Description

The function MLICwgee provides the model selection under the syntax of the wgeesel package.

Usage

MLICwgee(models, candidate.cor.sets, data, model_mis, family,r,id,time)

Value

A list with two items: model selection result based on ELCIC; An object of "wgee" based on the selected model.

Arguments

models

A list of formulas. See corresponding documentation to wgeesel.

candidate.cor.sets

A vector containing candidate correlation structures. It can be any subset of c("independence","exchangeable", "ar1").

data

A data frame containing the variables in both the main model and the missing model.

model_mis

A formula used in the missing data model.

family

A description of the error distribution and link function to be used in the model. The details are given under "Details".

r

A vector indicating the observation of data: 1 for observed records (both outcome and covariates are observed for a given subject), and 0 for unobserved records. The default setup is that all data are observed.

id

A vector indicating subject id.

time

The number of observations in total for each subject.

Details

Three commonly used distributions are considered: "gaussian", "poisson", "binomial". For the current package, the identity link is considered for a "gaussian" distribution; the log link is considered for a "poisson" distribution; the logit link is considered for a "binomial" distribution;n;

Examples

## tests
# load data
data(wgeesimdata)
family<-binomial()
r<-wgeesimdata$obs_ind
id<-wgeesimdata$id
time<-3
dat <- data.frame(y=wgeesimdata$y, wgeesimdata$x,x_mis1=wgeesimdata$x_mis[,2])
models <- list(y~x1+x2)
model_mis<-r~x_mis1
candidate.cor.sets<-c("exchangeable")
##not run
#output<-MLICwgee(models, candidate.cor.sets,data=dat,model_mis,family,r,id,time)
#output$model.selection
#output$wgee.output